4.5 KiB
Frequently Asked Questions
There's no great place to put these in our current files, but these sorts of questions come up enough, that we wanted to put them down somewhere to easily link responses.
Can I self advertise here?
Yes, provided it's FOSS, has a code of conduct (even if it's a solo project), and a No AI Policy, you can add your project as an alternative to any tainted software in which it is a good fit, and you can also open a Call for Collaborators Issue.
Do closed source or proprietary products belong on this list?
No, this repo is only for open source projects.
Can we split the "Permissive AI Policy" tag further?
No, beyond "vibecoded" being added in some extreme cases. We don't want to have this conversation continually, as it is subjective what "severity" something falls under. The decision to add these tags took us a long time, and unless you have another specific tag that cannot be covered under any other tag, we're currently not open to discussion about this. This repo is already quite a lot to manage.
Why are you targeting/attacking/harassing my favorite project?
We do not "target", "harass", or "attack" any projects. We, in fact, have a very strict policy on not doing that. This repo is one based on ethics and is intended as a call to action. If a project meets the definition of one of our tags, it will see itself added to this repo. It can take steps to stop doing whatever it is in that definition and we'll remove it. (See also the guide for removal.)
If a repo had slop before, isn't it permanently tainted? Aren't all forks of slop therefore also slop?
No. See also: guide for removal.
Eventually, the slop will likely either be removed, replaced, or deprecated.
We're also against AI for more than just code quality reasons. There's plenty of ethical reasons to be against continuing to use AI:
- environmental
- ties to the war industrial complex
- effect on hardware prices hurting newcomers to the industry (especially those already disenfranchised)
The commitment to stop using AI stops those above ethical issues. It's like going vegetarian. You're still a vegetarian if you previously consumed meat, but now you don't. You've still reduced suffering and helped the environment. A commitment to doing better is better than never stopping a harmful action.
Is a project tainted if it has dependencies that are tainted?
This one is complicated, but in short: No, unless the dependencies are literally AI libraries. If the dependency is, for example, the ChatGPT SDK or something similar, then yes, the project is tainted. If the project is just built with something like Python, which is also tainted, we do not automatically consider it tainted, as the language is a general purpose language and there are very few languages not already tainted or slated to be tainted right now. We take a harm reduction philosophy and we believe that, especially in the case of established projects, demanding that they switch the core language of their project is not a simple undertaking.
Can you advise us on the ethical usage of AI?
No, but you may find the work of DAIR interesting. Also, see our Why not LLMs section.
What about "Open Source AI"?
Yes, simply being open source does not resolve the issues listed in our Why Not LLMs? section and thus such projects will still be included on this list.
Why is removal harder than landing on the list in the first place?
Because the project betrayed our trust. Now they have to try harder to rebuild that by committing to doing better in an accountable, public, concreate, and enforcable way. If the project is committed to doing better by rejecting AI in the future, putting out a No AI policy or updating their contributing docs, shouldn't be that difficult.